Hi,

Good question. After I hit send, I realized that the output of cdi_inq or
inquire would be helpful.
(These outputs show that we don't have a binding issue)

CDI Output for the same tape drive:
crm(root):/home/rschuk>cdi_inq -f /dev/rmt/2cbn
Standard Inquiry data:
 Vendor:        IBM
 Product:       ULTRIUM-TD3
 Rev:           8711

VPD pages supported:
        Pages 00 80 83 87 b0 c1 e0 e1 ef
 Serial number page (80):
        7410810032
 Device ID page (83):
        ATNN=IBM     ULTRIUM-TD3     7410810032
Non-standard pages displayed only with -v parameter
  cdi_info.drivestat is:
    status = 1, DRIVE_STATUS_NO_ERROR
    msg = Drive reports no error - but state is unknown

Inquire output for the same tape drive:
[email protected]:IBM     ULTRIUM-TD3     8711|Tape, /dev/rmt/2cbn
                                           S/N: 7410810032
                                           ATNN=IBM     ULTRIUM-TD3
7410810032

I did confirm that we can mount and unmount the tape drive manually.
Actually, once it fails to mount the drive, I eject it using mt, and then
unmount it via the NMC GUI.
One other fact: These tape drives are on a Data Domain VTL. Probably not
relevant, but full disclosure and all that.

--Dave

-----Original Message-----
From: Stanley R. Horwitz [mailto:stan AT temple DOT edu]
Sent: Wednesday, February 06, 2013 4:35 PM
To: EMC NetWorker discussion; <dave2 AT CAMBRIDGECOMPUTER DOT COM>
Subject: Re: [Networker] serial number confusion

Dave,

It sounds like a binding problem. Are those tape drives connected via a
SAN switch? I suggest you check the connectivity of those drives and make
sure the connectivity is configured correctly at the OS level. What does
"inquire" say when you run it on the dedicated storage node?

On Feb 6, 2013, at 4:16 PM, <dave2 AT CAMBRIDGECOMPUTER DOT COM>
 wrote:

> Hi,
>
> We are seeing a case where NetWorker is failing to mount a tape on a
> new dedicated storage node due to it believing there is a serial
> number mismatch.
>
> The first time we saw this, I had configured the drives manually (via
> jbedit), so there was no serial number in the tape drive resource.
>
> The second time, I used dvdetect and the gui based autochanger
> configuration method.
>
> The server's daemon.log shows the following:
>
> .
>
> 38752 02/06/13 11:07:29  nsrd rd=crm:/dev/rmt/2cbn Eject operation in
> progress
>
> 42506 02/06/13 11:07:29  nsrd media alert: device "/dev/rmt/2cbn":
serial
> number                          mismatch, check system device ordering.
> Expected "Serial Numbers:ATNN=IBM     U
> LTRIUM-TD2     351A600203:351A600203", found ":ATNN=IBM     ULTRIUM-TD3
> 7410                         810032:7410810032"
>
> 72504 02/06/13 11:07:29  nsrd rd=crm:/dev/rmt/2cbn eject operation
failed:
> read                          open error: drive status is 65329 35 The
> drive serial number has changed
>
> 7224 02/06/13 11:07:30  nsrd [Jukebox `cdl', operation # 910]. read open
> error:                          drive status is 65329 35 The drive
serial
> number has changed
>
> .
>
> I'm noticing the following:
>
> 1.       the serial number mismatch says "/dev/rmt/2cbn," not
> "rd=crm:/dev/rmt/2cbn"
>
> 2.       The serial number of rd=crm:/dev/rmt/2cbn is 7410810032. This
is
> in changer "DD1"
>
> 3.       The serial number of /dev/rmt/2cbn (on the backup server) is
> 351A600203. This is in changer "CDL-old"
>
>
>
> Environment Info:
>
> NSR Server is Solaris, NetWorker 7.6.4.5
>
> NSR Dedicated Storage Node is Solaris, NetWorker 7.6.4.5
>
> Anyone see this, or have any thoughts about it?